Automotive companies look to the skies for clues on autonomous designs

While it’s far from a completely driverless system, there’s a reason why Tesla Motors named the semi-autonomous feature in its cars “autopilot.” Consumers may not be entirely sold on the idea of computers handling all aspects of driving, but airlines have effectively used automated flying systems for decades.

The push to develop self-driving cars could transform the auto industry from one that appeals to drivers with messages about design, handling, and performance, into a passenger-driven business. Several automotive suppliers say the aerospace industry offers clues on how to operate self-driving car fleets, so they’re working with airplane producers and studying sky-bound designs.

“With human-machine interaction in the airline industry, there’s a lot of knowledge on how a machine and a pilot work together to fly a plane,” says Erik Coelingh, technology advisor for Zenuity, a joint venture between automaker Volvo Cars and safety equipment supplier Autoliv, that is developing advanced safety and autonomous driving systems. “How do you make sure that the pilot really understands what’s happening with the plane? That’s an important consideration with semi-autonomous driving and autonomous driving. There’s going to be more human interaction with machines.”

Technology crossover between automotive and aerospace has been taking place since the births of both industries, but it has accelerated in recent years as each group faces new challenges. Automakers facing tighter fuel economy and emissions standards have learned a great deal about lightweighting from Boeing and Airbus, while the major aerospace providers have turned to automakers for expertise in boosting production as orders for new planes have outstripped production capacity.

With autonomous driving, lessons from plane producers could be invaluable, engineers say, not only in direct areas such as redundant safety systems and machine interfaces, but in the cultural shift that could be coming if cars in the future will be owned by airline-like fleets instead of individual consumers.

Airline seating, for cars

Adient, formerly the automotive seating division of Johnson Controls, is emblematic of auto industry’s work with aerospace producers. In January, Adient formed a joint venture with Boeing to produce airplane seats and seating systems. Dr. Detlef Juerss, vice president of engineering and CTO at Adient, says the work with Boeing will be critical to developing seats suitable for fully autonomous cars.

“New players keep coming into this market. Autonomous vehicles have new business models that will lead to completely different interiors,” Juerss says. “With autonomous vehicles, seats become more important. When people aren’t driving the vehicle, personal comfort will be more important than styling, handling, or power in differentiating products.”

Juerss adds that airlines have differentiated user experiences through seating since the advent of first class, and he sees the automotive industry following that lead.

“It’s going to be a range – everything from hard plastic, bench seats like you’d get on a commuter train to the pampered pods you get on an Emirates business class seat,” Juerss says.

Removing vehicle drivers also removes the traditional car instrument panel (IP), the main point of interaction between the human and the machine. Juerss says a self-driving car won’t need human input to control the vehicle itself, but passengers will want other functions currently housed in the IP – controls for heating and air conditioning, infotainment system controls, and data screens for GPS positioning and other traffic awareness information (just because you’re not the one driving doesn’t mean you won’t want to know if you’re almost at your destination).

“A lot of the convenience and features that have been in the instrument panel need to move to the seats – safety features, controls, infotainment,” Juerss says, adding that airplanes already have such controls integrated into armrests or seatbacks. “That all becomes more personalized and less a function of what the owner wants in an autonomous vehicle, so control design will be key.”

Juerss adds that automotive suppliers are finding concrete inspirations from aerospace designs. When designing car seats, engineers tend to place hard connection points for seatbelts and other safety features on the vehicle frame. So, when a crash occurs, the belt is attached to the strongest portion of the vehicle to keep passengers in place. Airplane seats are more modular and tend to integrate the belt attachments and the seatbelts into the seats, requiring stronger joints and tougher materials in some cases. Future automotive seats could take that integrated approach to enable a wider variety of seating options – seats placed around the inside diameter of the vehicle with passengers facing each other, for example.

The safety engineering task to enact that vision is a tough one, Juerss says. In a standard vehicle, the vast majority of crashes take place at the front or rear of the vehicle, so generations of safety systems have optimized safety for forces coming from those directions. Keeping a side-facing passenger in position in a frontal collision would be tougher. Adient engineers are working with Zenuity joint-venture partner Autoliv to integrate safety features into seats, he adds.

Redundant backup systems

While Adient’s interest in aerospace designs focuses more on comfort, Zenuity Technology Advisor Coelingh says his company is more interested in how aircraft makers deal with system safety. In many automotive applications, a failed sensor could lead to poor engine performance or loss of anti-lock braking systems (ABS), situations that rarely put drivers or passengers at risk because the human driver will be able to deal with it. In aerospace, system failures are unacceptable, requiring engineers to build backups for nearly every system, and oftentimes backups for the backups.

“Technology that we’re more familiar with in airplanes – redundancy, dual computers, dual software modules – that’s entering the automotive industry right now,” Coelingh says.

Autonomous vehicles are becoming available to the public in five stages, defined by SAE Int’l., ranging from assistance technologies to ones that completely manage all driving tasks. A few test fleets are pushing those outside boundaries today, but most vehicles haven’t progressed beyond advanced driver assistance systems (ADAS). Coelingh says that once the technology passes from being driver-supervised driving to no-human-involvement driving, the industry will near aerospace levels of redundancy.

“When you get to an unsupervised system, it is unreasonable to expect the driver to take over because he or she will be doing something else, and he will have no chance to intervene when something unexpected happens,” Coelingh says. “The autonomous system would have to be so robust that it could handle anything that could potentially happen on the road.”

He adds that he expects autonomous technology for personal vehicles to follow the traditional automotive technology route – a handful of very expensive cars will be fully autonomous soon, possibly within the next five years. Then, the technology will slowly migrate into mass-market cars – one feature at a time. For example, cars will be able to drive themselves in traffic jams or on the highway, when the amount of interaction needed to keep people safe will be minimal. The tipping point remains supervision – if a human driver is supervising all driving tasks, expensive redundant safety systems and massive sensor suites can be avoided. When driving becomes automated, costs increase.

At about the same time that expensive, luxury passenger vehicles hit the market, Coelingh expects equally pricey autonomous taxi-type vehicles to enter the market as well. Those early, expensive vehicles will drive costs down throughout time; technologies will improve, people will get used to having machines handle driving for them, and self-driving cars will become the norm.

“It will be many, many years from now before it becomes common to jump into a self-driving car. But the first steps will be taken within the next five years,” Coelingh says. “There are significant challenges, but we’re working through those. We still have far too many people dying every year in automobile accidents, and this technology can change that.”

Adient plc
www.adient.com

Boeing
www.boeing.com

Autoliv Inc.
www.autoliv.com

Volvo Cars
www.volvocars.com/us

Zenuity AB
www.zenuity.com

About the author: Robert Schoenberger is the editor of TMV and can be reached at 216.393.0271 or rschoenberger@gie.net.